Thought, Language, and Ontology, Essays in Memory of Hector-Neri Castañeda, ed. by W. Rapaport and F. Orilia (Kluwer, 1998): 3-24. THE GRAMMAR OF BELIEF Michael McKinsey Wayne State University W.V. Quine argued, notoriously and persuasively, that it is meaningless to quantify into opaque contexts. i I will refer to this claim as "Quine's thesis." The idea is this. Suppose that C is any sentential context containing an occurrence of a singular term such that substitution of coreferential terms (SI) is an invalid mode of inference with respect to this occurrence. Thus, in Quine's terminology, C is an "opaque" context. Now let C* result from C by substituting a variable v for the term at the relevant opaque occurrence. Then Quine's thesis says that binding of this occurrence of v by a quantifier outside the scope of C* results in a meaningless expres- sion. It has become common in recent years to endorse semantic theories of 'be-lieves' and other cognitive operators on which Quine's thesis is taken to be false. ii However, this tendency has not been accompanied by any direct argument against Quine's thesis as it applies to belief contexts. My aim in this paper is to provide such an argument. I will argue that, contrary to Quine's thesis, there are in fact meaningful sentences of English that involve quantification into opaque belief contexts. The basis of my argument will be the simple fact that there are meaningful sentences of English and other natural languages which in surface grammatical structure are indistinguishable from sentences that involve quantification into opaque contexts. Of course, on Quine's thesis, these sentences cannot really involve quantifying-in, since if they did, they would be meaning- less. I will examine a number of different hypotheses, most suggested by Quine himself, which if true would explain why, despite appearances, the sentences in question do not really involve